Document Type
Poster
Publication Date
Summer 2024
Abstract
LIGO, the Laser Interferometer Gravitational-Wave Observatory, uses advanced laser technology to detect gravitational waves, which are ripples in spacetime caused by massive cosmic events. Due to its extreme sensitivity, LIGO's detectors are affected by various sources of noise, ranging from minor temperature fluctuations to distant vibrations like passing cars. To address this, LIGO collects extensive data on the state of the detector through auxiliary sensors, such as thermometers. The auxiliary sensor data is reduced into feature sets, such as the strength and frequency of the signals in a given time window. This summer, I focused on developing machine learning algorithms (MLAs) which read patterns in the data found in these auxiliary features to predict glitches in the main gravitational-wave channel (strain data). This involved rebuilding data processing functions and unit tests for the TITAN pipeline, an MLA tool developed in Kenyon’s LIGO lab. By rewriting TITAN with classes, we created an updated version named TITANFALL. The TITANFALL pipeline, now debugged, has encountered an issue with one of the features in our datasets. Future work will involve constructing a new MLA for glitch-specific data frames and a more thorough investigation of the issues encountered when using the full feature dataset.
Recommended Citation
Temple, Josh; Chintala, David; and Wade, Madeline, "Developing and Refining Machine Learning Algorithms for LIGO's TITANFALL Pipeline" (2024). Kenyon Summer Science Scholars Program. Paper 694.
https://digital.kenyon.edu/summerscienceprogram/694